What it does
Gremlin's Chaos Engineering product lets engineering teams test how systems respond to faults in hosts, containers, services, and serverless workloads. The documented workflow installs and authenticates a Gremlin Agent, defines a service, adds health checks, and then runs a reliability test or fault-injection experiment. Experiments can reproduce conditions such as CPU pressure, packet loss, latency, disk pressure, or a host shutdown. Health checks can stop a test when the service becomes unhealthy, while results give teams evidence to improve their resilience.
Install, protect, and test a service
Gremlin's documented first task begins by installing its agent on the hosts, containers, or Kubernetes resources that will be tested, then authenticating the agent with the Gremlin control plane. Define the service being exercised and add health checks so the system's state is evaluated before, during, and after a test. If a health check becomes unhealthy, the documented reliability-test workflow halts the test and marks it as failed.
From there, choose a specific failure mode or scenario from the experiment library, such as network latency or packet loss, and begin with a small scope. Gremlin's product page describes automatic halt and rollback controls and recommends scaling experiments as confidence grows. A team can use the result to check whether failover, autoscaling, or dependency handling worked as intended; this is a test of the customer's own systems, not a monitoring dashboard that diagnoses an incident after the fact.
Gremlin's documentation advertises a 30-day free trial. The pages reviewed did not establish recurring prices. Because agents deliberately introduce faults, choose targets and health checks carefully before running tests against shared or production infrastructure.
Strengths
- The fault-injection library covers conditions such as CPU, memory, disk, network latency, packet loss, DNS, and shutdown.
- Health checks can halt tests when a service becomes unhealthy and provide a documented safety mechanism.
- The workflow supports hosts, containers, Kubernetes, serverless workloads, and on-premises environments.
- Documentation describes web-app, REST API, and CLI paths for viewing systems and running tests.
Watch out for
- The documented workflow requires installing and authenticating an agent on the systems being tested.
- Fault injection intentionally changes system conditions, so tests require deliberate target selection and health checks.
Best for
- SRE and platform teams testing service behavior under infrastructure or dependency failures
- Validating health checks, autoscaling, redundancy, or recovery procedures through controlled tests
- Teams organizing repeatable reliability experiments across hosts, containers, or Kubernetes
Sources
- Chaos Engineering | Gremlin (Oct 25, 2026)
- Overview > Gremlin Documentation | Gremlin Docs (Oct 25, 2026)
- Getting Started > Installing the Gremlin Agent | Gremlin Docs (Oct 25, 2026)